The invention relates to reflectometry systems in general, and more particularly to reflectometry systems and methods used to analyze change in a characteristic such as color or optical density of an area of a testing substrate. Such area may include colored spots that can represent a chemical condition in a specimen in a molecular recognition application (among other applications) to discern meaningful information from contrast signals.
It is known in the art to analyze biological and other specimens using optical techniques, including automatic optical analysis systems. Often it is desired to detect the presence of a particular analyte potentially within a specimen by testing for a reaction of the analyte with a specific reagent that can bind to the analyte. Such molecular recognition reactions represent the complexing of molecules that possess a high binding affinity to each other.
For example, consider the detection of the analyte human chorionic gonadotropin (HCG), a hormone present in the urine of pregnant women, as an indication of pregnancy. A few drops of urine are exposed to a substrate having a reagent thereon that is known to bind to HCG. In performing such testing, additional reagents may be added to the testing substrate such that a change in characteristic (e.g., color) in at least a portion of the results if HCG is present in the urine. The resultant color change can be clearly visible to the untrained eye and can serve as a home pregnancy test.
At best, the human eye can give qualitative results. But often tests do not produce readily ascertainable xe2x80x9cyesxe2x80x9d or xe2x80x9cnoxe2x80x9d results that are unambiguously apparent, even for a laboratory technician who is experienced in performing the tests and reading the results. Also, for some tests, quantitative results are desired, for example the measurement of progression or extent of a disease state. Visual examination of color and contrast changes are only qualitative, and the eyes of one observer will have a different sensitivity than the eyes of another observer. Even if two observers with identical visual sensitivity could be found, fatigue and subjective judgment differences could provide different results for identical data.
Many other molecular recognition applications, both immunological and non-immunological, can provide a meaningful contrast signal that can be analyzed using contrast data. In addition to immunological pair applications (e.g., antibody-antigen xe2x80x9cAb-Agxe2x80x9d), various sandwich format matrix assay techniques such as Ab/Ag/Ab, or Ag/Ab/Protein A-gold can generate meaningful contrast signals, as can bindings between avidin-biotin derivatives, or lectin-carbohydrate binding. Many applications use hormone receptors as molecular recognition sites, and reaction specific binding is a powerful analytical tool used in DNA hybridization. Such applications would benefit if more reliable and automated analytical tools could be provided.
Various systems have been attempted to provide an automated reliable system for analyzing results obtained from immunological and non-immunological molecular recognition applications. In some tests, especially those-involving immunoassay devices, it is necessary to discern the presence and reflectivity of dye-colored spots relative to the background area surrounding the spot. For example, when testing electrophoretic immunoblots (xe2x80x9cWestern Blotsxe2x80x9d) a densitometer is used to measure optical density of light reflected from nitrocellulose strips. But color density of antibody-produced color bands in the strips can vary, as can the background color. As a further complication, densitometers used in such tests cannot measure more than a single point in a color band. Thus, while densitometry can produce automated results, the results may vary greatly and can be highly inaccurate.
U.S. Pat. No. 5,006,464 to Chu et al. discloses the use of reflectometry to more rapidly quantitate the results of immunoassay tests. Suppose, for example, it is desired to examine human blood using such rapid immunoassay testing. A few drops of a blood, serum, or plasma specimen are deposited onto a testing substrate of an analytical device. The testing substrate is oftentimes a porous membrane that has one or more receptor chemicals bound thereon at discrete areas of the membrane that bindingly react to one or more target analytes, if present within the blood. Typically, after addition of the blood specimen, a few drops of a labeled reagent that may include a colored dye are added to the testing substrate. Finally, a few drops of a washing solution may be added to the testing substrate to remove any residual reagents that have not specifically bound to the discrete areas of the membrane where the receptor chemicals are located. The presence of the target analyte in the blood may then be indicated by the presence of a dye-colored spot on the device, relative to the uncolored surrounding background (which will be the area of the testing substrate that does not have receptor chemicals bound thereon). Such testing of course is not limited to the diagnosis of a particular analyte within blood, but may also be carried out with other types of specimens, biological or otherwise, that may contain target analytes of interest.
FIG. 1A depicts an exemplary device 10 that may be used to carry out the above-described Chu type analysis. Device 10 may measure perhaps 1 cm square and, for reasons of economy, is fabricated from several layers of cardboard (or sometimes plastic) including upper and lower layers 20, 30. Upper layer 20 defines an opening 40, perhaps 8 mm in diameter, that exposes a surface 50 that lies higher than bottom layer 30. Surface 50 defines a membrane (or substrate) whose preferably porous surface contains at least one immobilized receptor chemical that will cause a binding reaction with a target analyte, when present in a specimen.
FIG. 1B, is a top-view of the device depicted in FIG. 1A, after the above-described testing procedure has been carried out. Spot 60, where a receptor chemical is adhered onto the testing substrate, is shown as being rather (ideally) circular and having a color that is in sharp contrast to the surrounding upper surface 70 of device 10, indicating that the analyte of interest was present in the specimen tested. The challenge is to determine reflectivity of spot 60 relative to surrounding region 70. Stated differently, the challenge is to distinguish between signal from spot 60 and a background reference level from region 70. It will be appreciated that a small change in either signal can result in a substantial change in the difference between the two signals.
The ability to differentiate reflectivity signal from the background reference level permits one to arrive at a meaningful conclusion as to the presence or absence of the target analyte in the specimen. For example, FIG. 1C shows device 10 with no spot whatsoever, e.g., no binding reaction has occurred, and the target analyte is absent from the specimen. In FIG. 1D, a dark spot is present, but as may often be the case, the spot is not uniform in shape and may be smaller in size than anticipated. Spot size can be affected by the area of membrane 50 that was originally impregnated or otherwise treated with chemicals before the specimen was introduced. FIG. 1E depicts a uniform spot 60, but of less opaqueness than the spot shown in FIG. 1B, while FIG. 1F shows a non-uniformly shaped spot of less opaqueness than was shown in FIG. 1A or FIG. 1C.
It can be rather difficult to accurately discern reflectivity of spot 60 relative to the surrounding region 70. This is especially true if distinguishing signal from noise is to be accomplished rapidly, preferably in an automated fashion, without requiring trained personnel. In many applications, the difference between a positive reading and a negative reading can be less than about a xc2x11% change in reflectivity. In practice, borderline readings often occur when spot reflectivity is perhaps 98% of background reflectivity. In practice, a reading of 98.4% would be negative, while a reading of 97.6% would be positive. Thus, it is important that reflectivity be interpreted accurately and consistently if meaningful data are to be obtained. Even a trained human eye cannot resolve changes in reflectivity as small as a few percent. The human eye is also a poor reflectometry instrument when one must examine many devices 10 within a given time. Fatigue, subjective differences in observation, and errors can result in different analysis results, even when reflectivity changes are large enough to permit differentiation with the eye.
Accordingly, an automated system for correctly ascertaining reflectivity data is required. Typically after exposure to a specimen, and post-exposure treatment and dyeing as described above, device 10 is inserted into a reflectometry device. A reflectometer subjects at least the spot area of the device to light, and attempts to measure intensity of light reflected from the spot area (the signal) and from representative surrounding region (the reference signal). But unless the device is inserted into the reflectometer in an aligned manner, light from the receptor spot area may not be accurately measured. Indeed, for low density spots, it can be difficult to discern dot intensity from background area intensity. The colored spot itself may not be uniformly colored, and the spot signal reading may be highly dependent upon what portion of the receptor area spot was examined for reflected light intensity.
Further, conventional reflectometer signal processing units tend to be relatively large and cost upwards of $1,500 per unit to manufacture, exclusive of a signal processing computer, typically requiring at least a 80486-type microprocessor. For example, charge control devices (xe2x80x9cCCDsxe2x80x9d) not unlike what is found in modern video cameras were often used to try to measure intensity. However pixel-to-pixel sensitivity variations in such devices made calibration and consistent data readings somewhat difficult. Further, in addition to bulk and expense, such systems were not especially robust mechanically.
Even if the above problems did not exist, the surface of membrane 50 commonly provides an irregular surface topography. This irregularity is present with inexpensively produced devices containing cardboard or paper membrane surfaces, as well as with higher quality woven cloth surfaces. As used herein, the term membrane is understood to include paper, cardboard, cloth, and other commonly used membrane materials used in producing a device 10. Although optical-quality membranes (e.g., glass, ceramic, metal) would present a substantially planar surface, these more expensive materials are not suitable to hold the desired chemicals, while allowing for diffusion of water and the like.
As a result, even if color intensity in the receptor area were somehow completely uniform and very discernable from the background region, irregularities in the membrane surface would still introduce error into the intensity readings. Although for immunoassay tests in which a high intensity spot results, a 1%-2% intensity readout error might be acceptable, for other tests an error of 1% or 2% may result in a completely false result.
What is needed is a technique for rapidly and accurately analyzing intensity signals produced by spots on devices, where spot density contains important data. Such technique should be applicable for immunological and non-immunological molecular recognition applications. Such technique should exhibit improved signal to noise characteristics, and preferably should compensate for irregularities in the device topology. The technique should be mechanically robust, simple enough for a layman to operate, and should be relatively inexpensive to produce and maintain.
The present invention provides such a method and a system for carrying out the method.
The present invention provides a self-contained system that examines a device containing a membrane upon which is formed a dyed spot surrounded by undyed background. The dyed spot may be the result of molecular recognition, including immunologically and non-immunologically based applications, among other applications. The invention uses light reflectivity to determine relative spot signal to background reference signal, to glean meaningful and information as to spot intensity. This information allows quantifying a target analyte or chemical or substance, whose presence contributed to the spot dye characteristic.
The membrane surface will include a spot dyed region that is responsive to the presence of the target substance, which spot region is typically surrounded by an undyed background region. In a preferred embodiment, the present invention alternatively illuminates the spot region and the background region, and measures intensity of light reflected from each region. In another embodiment, the membrane surface is illuminated (dyed spot region and background region) and then the spot region or the background region but not the spot region is illuminated. The intensity of light reflected during the various modes of illumination may be subtracted to cancel the zero output drift effects associated with amplified signals from the light detectors in the absence of light. The emitter light may come from preferably three substantially identical light sources, e.g., light emitting diodes (xe2x80x9cLEDsxe2x80x9d), one of which is aimed at the spot region and two of which air aimed at the background regions. A master clock unit actively drives the LEDs to alternately illuminate the spot and the background. Alternatively, the master clock unit can cause the LEDs to simultaneously illuminate spot and background regions, alternating with time periods during which only the spot or only the background without the spot is periodically illuminated. Other embodiments provide a single light source, whose light output is directed uniformly upon spot and background regions, for example using a diffusive or reflective reflector unit. In such embodiment, control over what device surface regions receive light is controlled by a master clock unit controlled perforated rotating mechanical shutter, or by a solid state preferably liquid crystal shutter. In the various embodiments, the light source(s) define one plane and the membrane containing the spot and background regions define a second plane. The light source(s) may be disposed at either plane, and the light detectors disposed at the remaining plane. The emitted light spectra may be visible, invisible, infrared, ultra-violet, X-ray, and may be produced from a variety of sources including LEDs, laser diodes, incandescent lights, among other light sources.
The reflected light, which is to say a fraction of the emitted light, is then detected preferably by photodetectors (xe2x80x9cPDsxe2x80x9d) spaced-apart a chosen multiple of 90xc2x0 azimuthally. Depending upon the topology of the device membrane surface, there preferably will be two PDs that are azimuthally spaced-apart 180xc2x0 or 90xc2x0.
The invention provides two sub-systems: a sub-system that substantially reduces skew error due to uneven device topography, and a phase lock-in amplifier sub-system to enhance detected signal/noise by measuring signal voltage without producing noise.
In the first sub-system, the present invention recognizes that poor topography associated with commonly used membranes can contribute an uncertainty error of at least xc2x11% in reflected light intensity readings. This error, which is approximately sinusoidal, is substantially cancelled by monitoring reflected light intensity using two PDs spaced-apart from each other 180xc2x0 azimuthally. Should a membrane be fabricated from cross-linked cardboard, then the two PDs are spaced-apart from each other 90xc2x0 azimuthally. So disposing a pair of PDs a chosen multiple of 90xc2x0 azimuthally apart, and averaging their output signals can improve accuracy xc2x11% or more.
In the second sub-system, the averaged PD output signal is simultaneously positively and negatively amplified, and the positive and negative amplified signals are sampled with a switch that operates synchronously with the active region of the LED drive signals. The result is that only in-phase signals occurring during the active LED drive signal portion can be sampled. Further, signals that are present at the output of the PD summing amplifier are essentially doubled in effective amplitude, thereby enhancing signal/noise ratios.
The output of the switch is coupled to a lowpass filter that attenuates frequencies higher than the measurement frequency, perhaps 3 Hz. The measurement frequency is substantially lower than the LED clock switching frequency. As such, the lowpass filter output is substantially free of noise, other than noise components within a few Hz of the clock frequency. The lowpass filter thus recovers a signal proportional to spot light intensity, and exhibits excellent signal to noise characteristics.
Other features and advantages of the invention will appear from the following description in which the preferred embodiments have been set forth in detail, in conjunction with the accompanying drawings.